Can fatigue cracks form under compression-compression stress cycles? 1

[bugbus]

I asked this question on the Metal & Metallurgy forum but did not have much luck.

I have heard conflicting answers to this question before, and I have to say that my background is not in materials but in structural engineering. So my experience with fatigue is less on the theoretical side and more on the code-based side.

Supposing that a steel element is subjected to purely compressive load cycles (i.e. min and max stresses are both compressive), and that there are no residual stresses of any kind in the element (such from welding, hot rolling, plastic deformation, or whatever), is it possible for fatigue cracks to form? I have heard that the answer is "yes", but that the rate is in the order of 100x slower than in tension. However, I have no way of knowing if this is right.

Is this any different, then, to a situation where the stress cycles are in tension, but due to residual compressive stress in the element (such as from peening, or some other effect), the overall stress experienced is still always compressive?

And I suppose the opposite situation is where there is a residual tensile stress in the element, but the stress cycles are purely compressive. At the point under consideration, at least part of the stress cycle occurs in tension. Is it possible for fatigue cracks to form in this situation?

 

[rather_be_riding]

gusmurr,

I'll start by professing a greater than usual degree of ignorance in this field. I'll be interested to see responses from other, more specifically qualified parties.

[ul]
[li]I believe it is entirely possible. If you load the material past yield, you could end up with tensile residual stresses during periods of minimal compressive loading. Therefore the crack growth will actually occur each time you unload rather than each time you load if that makes some kind of sense.[/li]
[li]In the situation where you've retained compressive stress, you should extend your fatigue life as you may never or rarely experience tensile stresses in the part.[/li]
[li]In the last scenario, you could well experience crack growth as your stress will vary from tensile through compressive and back again each loading cycle. This is essentially the same as the first scenario post yielding.[/li]
[/ul]


Of course, truly purely compressive loading isn't always all that common anyway so it's plausible that a seemingly compressive load results in shear and tensile stresses somewhere in a component.

 

[desertfox]

I have little knowledge in this area but I have always been lead to believe that fatigue cracks grow under tensile stresses and not compressive stresses.
If you have tensile residual stresses in a component which is then only subject to compressive load, I would concur with other posters that fatigue cracks could still occur.

“Do not worry about your problems with mathematics, I assure you mine are far greater.” Albert Einstein

 

[BJI]

Supposing that a steel element is subjected to purely compressive load cycles (i.e. min and max stresses are both compressive), and that there are no residual stresses of any kind in the element (such from welding, hot rolling, plastic deformation, or whatever), is it possible for fatigue cracks to form? I have heard that the answer is "yes", but that the rate is in the order of 100x slower than in tension. However, I have no way of knowing if this is right.

I would say technically the answer is yes, but for any real application it would be no. There are always small defects which can lead to singularity stresses and micro plasticity. Over sufficient number of cycles you could get micro-void coalescence and therefore crack nucleation.

Is this any different, then, to a situation where the stress cycles are in tension, but due to residual compressive stress in the element (such as from peening, or some other effect), the overall stress experienced is still always compressive?

Beneficial compressive stresses, such as that from peening, would essentially blunt the peak but your stress range would remain the same. This gives an improvement in fatigue life but is quite different to the first scenario.

And I suppose the opposite situation is where there is a residual tensile stress in the element, but the stress cycles are purely compressive. At the point under consideration, at least part of the stress cycle occurs in tension. Is it possible for fatigue cracks to form in this situation?

This is quite a common situation. Often weld residual stresses are at yield levels (without any PWHT), therefore even though your loading is compressive, the stress cycles are tensile.

Here is a previous reply to your related question, is this the same application?

https://www.eng-tips.com/viewthread.cfm?qid=478363

 

[rb1957]

I think it would depend on whether you caused plasticity (and so residual tension stresses would appear in the structure).

another day in paradise, or is paradise one day closer ?

 

[bugbus]

Thanks everyone, seems like there is a bit of a consensus that as long as there is residual tensile stress, cracks can form.

BJI, I had completely forgotten I had asked a similar question before, thanks for the old link.

 

[Settingsun]

Gusmurr, I don't know the answer but the commentary to AS4100 says to consider compression stress range to be as damaging as tension unless it can be shown to be otherwise. C11.3.2.

 

[rb1957]

I understood that shot peening expanded the surface creating a compression field on the surface (and a smaller sub-surface tension reaction). And since most fatigue cracks initiate on the surface there's an improvement in fatigue life. I would add the surface stress from shot peening to the applied stresses, if shot peen induced stress is higher than the cyclic maximum then, yes, you'd've have (at least on the surface) compression-compression cycles.

another day in paradise, or is paradise one day closer ?

 

[bugbus]

steveh49, I'm not sure if it is the same in AS 4100, but in AS 5100.6, they say that the compression part of the stress cycle can be reduced by 40% provided the element is not welded or is stress-relieved. I think that is a bit of a clue that this is to do with residual stresses.

 

[WARose]

Supposing that a steel element is subjected to purely compressive load cycles (i.e. min and max stresses are both compressive), and that there are no residual stresses of any kind in the element (such from welding, hot rolling, plastic deformation, or whatever), is it possible for fatigue cracks to form?

I'd say "yes" for the simple reason that even with "pure" compressive stresses.....you will still have (for all practical purposes, in most shapes) concentrations and shear stresses. So it will never be "purely" anything.

 

[Ron]

The simple answer is Yes.